Fragment of the 11S Proteasome Activator Complex, Reg - American

May 10, 2007 - facilitate biomarker hunting and validation which can be named ... MALDI imaging • Ovarian cancer • biomarker hunting • biomarker...
0 downloads 0 Views 1MB Size
Download PDF
Specific MALDI Imaging and Profiling for Biomarker Hunting and Validation: Fragment of the 11S Proteasome Activator Complex, Reg Alpha Fragment, Is a New Potential Ovary Cancer Biomarker Remi Lemaire,†,# Sonia Ait Menguellet,†,‡,# Jonathan Stauber,†,# Valerie Marchaudon,†,‡ Jean-Philippe Lucot,‡ Pierre Collinet,‡ Marie-Odile Farine,§ Denis Vinatier,‡ Robert Day,| Patrick Ducoroy,O Michel Salzet,*,† and Isabelle Fournier*,† Laboratoire de Neuroimmunologie des Anne´lides, FRE-CNRS 2933, MALDI Imaging Team, Cite´ Scientifique, Universite´ des Sciences et Technologies de Lille, 59650 Villeneuve d’Ascq, France, Clinique de Gyne´cologie, Hoˆpital Jeanne De Flandre, CHRU Lille, 59000 Lille, France, service d’Anatomie et de Cytologie Pathologiques, Hoˆpital Calmette CHRU, 59037 Lille Cedex, France, De´partement de Pharmacologie, Faculte´ de me´decine, Universite´ de Sherbrooke, Sherbrooke, Que´bec J1H 5N4, Canada, and Plateforme Prote´omique, IFR 100-Sante´-STIC, 8 Boulevard du Mare´chal de Lattre de Tassigny, 21000 Dijon, France Received May 10, 2007

MALDI imaging mass spectrometry represents a new analytical tool to directly provide the spatial distribution and relative abundance of proteins in tissue. Twenty-five ovary carcinomas (stages III and IV) and 23 benign ovaries were directly analyzed using MALDI-TOF MS. The biomarker with the major prevalence (80%) has been fully identified using MALDI MS and nanoESI MS and MS/MS after separation by RP-HPLC and trypsin enzymatic digestion. This marker with an m/z of 9744 corresponds to 84 amino acid residues from the 11S proteasome activator complex, named PA28 or Reg-alpha. Validation of this marker has been performed using MALDI imaging, classical immunocytochemistry with an antibody raised against the C-terminal part of the protein, specific MALDI imaging, and Western blot analysis. The validation, using immunocytochemistry, confirmed the epithelial localization of this fragment with nucleus localization in benign epithelial cells and a cytoplasmic localization in carcinoma cells. This indicates that this antibody could be used to discriminate the borderline tumor cases. At this point, a multicentric study needs to be conducted in order to clearly establish the potential of this biomarker. Taken together these studies reflect that direct tissue analysis and specific MALDI imaging strategies facilitate biomarker hunting and validation which can be named pathological proteomics. Keywords: MALDI direct analysis • MALDI imaging • Ovarian cancer • biomarker hunting • biomarker validation • Proteasome 11S

Introduction Ovarian cancer is now the fourth leading cause of cancer death among women in the Europe and the United States, with more than 22 000 new cases and 16 000 deaths reported each year in the U.S.1 Despite advances in therapeutic management during the past decade, ovarian cancer remains the most lethal gynecological malignancy. Indeed, early stage disease is usually asymptomatic, and ovarian cancer typically presents with advanced disease, where the prognosis is poor. Ovarian cancer is diagnosed at a metastatic stage in more than 80% of cases and is associated with a 5-year survival of 35% in this popula* To whom correspondence should be addressed. E-mails: Michel Salzet ([email protected]) or Isabelle Fournier (isabelle.fournier@ univ-lille1.fr). † Universite´ des Sciences et Technologies de Lille. ‡ Hoˆpital Jeanne De Flandre. # Equal contribution. § Hoˆpital Calmette CHRU. | Universite´ de Sherbrooke. O Plateforme Prote´omique. 10.1021/pr0702722 CCC: $37.00

 2007 American Chemical Society

tion. In contrast, the 5-year survival for patients with early stage of the disease exceeds 90%. Screening ovarian cancer at the early stages is today the only way to significantly decrease mortality, without the need to change therapeutic modalities. Several strategies for diagnosis have been tested, in particular, endovaginal ultrasound and tumor biomarkers. However, due to their low sensitivity at the early stage of the disease, none of these tests have been retained for ovarian cancer screening in the case of high cohort. Among the biomarkers, CA 125 is one of the most studied. CA 125 has a sensitivity of 80% and a specificity of 97% in epithelial cancer (stage III or IV), but the sensitivity is around 30% in stage I and its increase can be linked to several physiological phenomena and also detected in benign situations.2 CA 125 is very useful in the case of risk populations’ diagnosis or to follow the illness evolution during the therapeutic treatment. In this context, CA 125 is insufficient as a single biomarker for ovarian cancer diagnosis. This opens the door to discover other biomarkers using a proteomic strategy.3-16 Journal of Proteome Research 2007, 6, 4127-4134

4127

Published on Web 10/16/2007

research articles At this time, two strategies have been undertaken by researchers. In fact, several groups have tried to identify ovarian cancer markers in plasma or serum using SELDI-TOF profiling or chromatography coupled with mass spectrometry.3,17-23 Other groups have developed a classical proteomic strategy using comparative 2D-gels followed by mass spectrometry.15,24-26 Here, we proposed another strategy based on peptide profiling by direct MALDI analysis of the tissues and MALDI imaging comparing carcinoma to benign tumors. With this strategy, several peptides were observed to only be present in carcinoma compared to benign samples with a statistically significant score. The most prevalent peptide was identified as a fragment of immunoproteasome PA28 after peptide extraction from tissues, separation by RP-HPLC, and MS analysis after trypsin digestion of the protein based on the peptide mass fingerprint and peptide sequence tag. This potential biomarker was then validated by immunicytochemistry (ICC), Western blot, and specific MALDI imaging experiments.29

Materials and Methods Samples. Tissues, ascites, and cyst fluids were obtained, with informed consent and institutional review board approval (CCPPRBM Lille: CP 05/83), from patients undergoing any ovarian tumor resection at Hospital Jeanne de Flandre (Lille, France). In total, 48 tissue samples from 25 grade III and IV ovarian cancer patients and 23 benign tumors were analyzed. Patient information was collected, including gender, age, treatment received before and after surgery, extent of surgery, current status (alive, alive with progressive disease, deceased, and cause of death), and survival from the time of original pathologic diagnosis. Samples were collected at the time of surgery, immediately frozen, and stored at -80 °C until analysis. Typically, 10-12 µm thick sections were cut using a cryostat and thaw-mounted on a flat electrically conductive sample slides. Histopathologic diagnoses were made by an anatomopathologist, blinded to the original clinical diagnosis, from subsequent H&E-stained sections. Ascites and cyst fluids were collected in heparin tubes and immediately subjected to an acidic extraction with 1 M HCl (v/v). Peptides Extractions from Tissues. Biopsies were frozen in liquid nitrogen before homogenization and extraction with 1 M HCl (w/5v), and incubated overnight at 4 °C with rocking. After centrifugation at 12 000 rpm for 30 min at 4 °C, the pellet was re-extracted once. The two supernatants were combined and loaded on 200 mg Sep-Pak C18 cartridges (500 µL extract/ cartridge; Waters). After a wash step with 5 mL of acidified water (0.05% TFA, Pierce), elution was performed with 5 mL of 60% acetonitrile in acidified water (0.05% TFA). The 60% eluted fraction was reduced in a vacuum centrifuge (Savant), mixed with 100 µL of acidified water (0.05% TFA), and fractionated on a C18 Revered-Phase HPLC column (4.6 mm × 25 cm, Interchim) equilibrated with acidified water (0.05% TFA). Elution was performed with a linear gradient of acetonitrile in acidified water (0.05%) from 0 to 70% at a flow rate of 500 µL/ min. Each fraction was collected manually before evaporated in SpeedVac vacuum and mixed with 50 µL of HPLC grade water. Each fraction was analyzed using MALDI-TOF MS before digestion with trypsin. Trypsin Digestion. After drying, samples (extracted peptides) were placed on ice for 30 min in 50 µL of protease solution (sequence grade-modified trypsin, Promega, at 0.02 mg/mL in 4128

Journal of Proteome Research • Vol. 6, No. 11, 2007

Lemaire et al.

25 mM ammonium bicarbonate). Digestion was performed overnight at 37 °C. Peptide extraction was performed twice for 15 min with 50% acetonitrile and 1% TFA for further MALDI MS analysis, or with 50% acetonitrile and 1% formic acid for further ESI MS/MS analysis. Trypsin digests were then lyophilized in a SpeedVac concentrator and resuspended in 5 µL of 0.1% formic acid. Material. R-Cyano-4-hydroxycinnamic acid (HCCA), sinapinic acid (SA), ammonium bicarbonate, trisma base, xylene, ethanol, Angiotensin II, Des-Arg-Bradykinin, Substance P, ACTH 18-39, ACTH 7-38, and bovine Insulin were obtained from Sigma-Aldrich and used without any further purification. Trypsin was from Promega. AspN and LysC enzymes were from Roche. Trifluoroacetic acid (TFA) was purchased from Applied Biosystems. Acetonitrile (ACN) p.a. and methanol p.a. were from J. T. Baker. Mass Spectrometry. 1. MALDI-MS Direct Analysis of Biopsies. Spectra were acquired on a Voyager-DE STR mass spectrometer (Applied Biosystems, Framingham, MA) with delayed extraction (DE) and a 337 nm pulsed nitrogen laser. HCCA was used as matrix at a concentration of 10 mg/mL in ACN-0.1% TFA/H2O (2:1, v/v). Matrixes were applied onto the tissue using a micropipette (typically 20 µL for a whole ovary slice) and then dried at room temperature. External calibration was performed using a mixed solution of peptides (bradykinin 1.6 µM, Substance P 1.6 µM, ACTH 18-39 1.6 µM, ACTH 7-38 3.2 µM, bovine Insulin 4.8 µM, and bovine Ubiquitin 4.8 µM in H2O). Sections were visualized in the mass spectrometer using a color CCD camera (SONY). Each spectrum is the average of 200 laser shots of the area of interest. Acquisition parameters were set as follow: acceleration voltage, 25kV; first grid voltage, 94%; guide-wire voltage, 0.05%; extraction delay time, 200 ns. 2. NanoESI-q-TOF MS and MS/MS Analysis. After desalting the trypsin digests on a C18 Zip-Tip (Millipore, Bedford, MA), they were loaded into nanoESI capillaries (Protana, Odense, Denmark). The capillaries were inserted into an Applied Biosystems Q-STAR Pulsar (Q-TOFMS) using an ion spray source. Doubly charged peptides were selected, fragmented by N2 collision, and analyzed by nanoESI-q-TOF. Protein identification was performed with MASCOT sequence query search program using Swiss-Prot database filtered for the taxonomy “human”. A tolerance of 1.2 Da for peptide in MS and 0.8 Da for MS/MS was set. Only protein sequences with MOWSE score higher than 32 (indicating significant homology or identity) and identified in several samples representing 3 significant MS/MS were considered. Methionine oxidation was defined as variable modification. MALDI Imaging and Specific MALDI Imaging. 1. MALDI Imaging. Frozen ovarian sections of 10 µm were obtained using a cryostat and immediately transferred onto a conductive Indium-Tin Oxide (ITO) glass (Bruker Daltonics, Wissenbourg, France). After the sections had dried for 5 min at room temperature, tissues were heated at 80 °C during 20 s for good adherence onto the slides. They were then rinsed in chloroform and directly analyzed using MALDI MS after covering the tissue section with matrix using a pneumatic TLC sprayer. For imaging experiments, sinapinic acid acid was used as matrix at a concentration of 20 mg/mL in 0.1% TFA in water/ACN (3: 7, v/v). 2. MALDI Specific Imaging. For specific MALDI imaging studies, ICC experiments were performed using classical protocols and were as follow. Ovary sections were incubated at room temperature with 500 µL of buffer (0.1 M PBS/1% BSA/

research articles

MALDI Imaging/Profiling for Biomarker Hunting and Validation

1% normal goat serum/0.05% Triton X100) for 30 min.27,28 The same buffer was used to dilute Reg alpha antibody (1/100) (Zymed Laboratories, Invitrogen, ref. 38-2400), and incubation was performed overnight at 4 °C. After 3 washes with PBS (0.5 M, pH 7.4), sections were incubated with anti-rabbit IgG photocleavable tagged secondary antibody (dilution 1/100) (Eurogentec29) overnight at 4 °C. The tissues were then rinsed 3 times during 5 min with ultrapure water to remove salts, and sections were kept drying at room temperature before matrix application. For specific imaging, both classical HCCA (10 mg/ mL 0.1% TFA in water/ACN (3:7, v/v) and solid ionic matrix HCCA/aniline (10 mg/mL 0.1% TFA in water/ACN (3:7, v/v) were used. Matrixes were applied using 20 µL of solution, deposited directly using a micropipette, and allowed to dry at room temperature. Tissues were then compared using microscopy. Acquisition was performed on an Ultraflex II MALDITOF/TOF (Bruker Daltoniks, Bremen, Germany) instrument equipped with a smartbeam laser at a repetition rate of 100 Hz. For image reconstruction, the FlexImaging v. 2.0 software (Bruker Daltonics, Bremen, Germany) was used. For MALDI imaging, 37 699 points covering the whole slice with 300 laser shots per position with masses ranged between 2000 and 20 000 were performed. For specific imaging, 2000 points were done with 100 laser shots per position. From each position, the software measures an average mass spectrum with its coordinates on the slice. Immunocytochemistry (ICC). Immunohistochemistry was performed from paraffin-embedded ovarian tissues, using a standard peroxydase-based staining method. Paraffin-embedded tissue sections (4 µm) were performed on a microtome and then dried for 12 h at 60 °C, dewaxed with xylene 100% for 5 min, and rehydrated in brief successive baths of alcohol 100%, alcohol 95%, and distilled water. Tissue sections were stored in distilled water for the following steps. Endogenous peroxydase activity was quenched by incubating the section in 10% H2O2 in distilled water for 5 min. Tissue sections were incubated with the primary anti-PA 28 Alpha C-term antibody (Zymed Laboratories, Invitrogen, ref. 3-2400) at a dilution of 1/50 in TBS (50 mM, pH 7.4) for 1 h at room temperature. Tissue sections were washed with TBS solution and were incubated successively with the secondary antibody (BiotinStreptavidine-Peroxydase, Horseradish Peroxidase (HRP) Conjugate) and the chromogen (3.3′-diaminobenzidine) according to the manufacturer’s instructions (Streptavidin-HRP, Southern Biotechnology and Associates, Inc.). Cellular nucleus were then colored by counterstaining the sections with hematoxylin. Western Blot. Western Blot analysis was carried out from frozen ovarian tissues. Protein samples were obtained by homogenization on ice of 100 µg of tumor tissue for 3 min in 1 mL of extraction buffer (50 mM Tris-HCl, 150 mM NaCl, and 0.02% sodium azide) supplemented with protease inhibitors. After homogenization on ice with an ultrasonic probe (3 successive steps of 30 s each), the mixture was placed under agitation at 4 °C for 1 h, followed by centrifugation at 13 000g at 4 °C for 30 min. The supernatant was removed and stored on ice for the following steps. To achieve protein denaturation before analysis, a 4% SDS buffer was added 1:1 to the supernatant, and then the mixture was warmed for 5 min at 100 °C. For Western blot analyses, 18 µg of proteins (measured according to the Bradford method) was applied onto a 3.912% Tris-Glycine-SDS-polyacrylamide gel (protean II minigel, system, Bio-Rad, France). Protein migration was achieved in a Tris-Glycine-SDS buffer (2 mM/129 mM/0.1w/v, pH 8.3) by

application of a tension of 80 V for 15 min followed by 180 V for 1 h. Gels were subsequently transferred onto a nitrocellulose membrane previously equilibrated for 30 min in transfer buffer refrigerated at 10 °C. Blots were blocked in 5% milk in PBS for 45 min and incubated overnight at 4 °C with a rabbit anti-PA 28 alpha C-term antibody diluted 1/250. Detection was performed by enhanced chimioluminescence (Amersham, France) after 1 h incubation with a peroxydase-conjugated secondary antibody (Aventis, Sanofi Pasteur, France). Statistical Data Analysis. Symbolic discriminant analysis (SDA) was used to analyze the protein profiles. SDA allows determining discriminatory signals and builds functions using these signals that distinguish sample populations based on their classification. Peak lists from MALDI spectra obtained with Flex analysis 2.4 and Flex control 2.0 input into the analysis program were clustered according to similarity, using a presence/ absence criterion. The Mann-Whitney U test and Spearman correlation test were used to test for differences in the classified staining levels of PA28-R expression between groups for immunocytochemistry. The values of Western blot analysis were estimated by densitometry and compared by ANOVA. Significant differences between groups were determined by Student’s t test. A P-value of less than 0.01 was considered to indicate statistical significance. All statistical analyses were performed on a personal computer with the statistical package StatView 5.0 (SAS Institute, Inc., Cary, NC).

Results and Discussion Mass Spectrometric Profiling of Human Ovarian Tumors. A total o 48 tissue samples (25 carcinoma and 23 benign tumor cell populations) was collected and analyzed by MS. Tissue sections were coated with matrix droplets anddirectly analyzed using MALDI MS serial sections which were collected and stained with Haematoxylin Eosin for histopathology. Over 500 individual mass spectra, representing benign cell populations from benign patients or cancer cell populations from ovarian carcinoma patients, were used for comparative analysis. The spectra were processed, and multiple spectra were averaged to generate one peak list per patient or tissue sample. At 100, individual peptide signals in the m/z range of 500-20 000 were detected (Figure 1a). The presence or absence of biomarkers between cancer and benign tissue as well as in ascite or in cyst fluid was research in this study (Figure 1b). In these conditions, we detected a putative biomarker (m/z of 9744) with a prevalence of 80% (Figure 1c) with a specific localization determined with MALDI IMS (Figure 1d). Characterization of the Ovarian Tumor Marker. Peptide extraction with HCl 1 M, followed by Sep-pack prepurification and HPLC separation on a C18 reversed-phase column, was performed. Each of the 60 fractions was then analyzed using MALDI-TOF. The MALDI analysis reveals that the fraction which eluted at a retention time of 46 min is the one containing the polypeptide at the expected m/z of 9744. In this fraction, the m/z 9744 peptide is more abundant. Thus, the fraction was subjected to trypsin digestion and analyzed both using MALDITOF (Figures 2a,b) and nanoESI-q-TOF for peptide mass fingerprint (Figure 2b,c). Trypsic digestion profiles obtained with MALDI, after removing the ions from trypsin autolysis, were used for databank interrogations (Figure 2a). Thus, ions detected with m/z of 536.25, 700.68, 787.7, 841.63, 890.57, 970.47, 1032.17, 1158.55, 1349.16, 1501.03, 1528.99, 1790.18, 1833.07, 2104.58, 2210.47, and 2280.02 gave a MASCOT top score of 54 for 1AVOB (11s regulator, chain B-human N; Journal of Proteome Research • Vol. 6, No. 11, 2007 4129

research articles

Lemaire et al.

Figure 1. (a) MALDI mass spectrometry profile of ovary carcinoma vs benign tumor (arrows indicate specific marker, in light blue are the common molecules identified in both tissues). (b) MALDI mass spectrometry profile of ascite fluid from ovary carcinoma vs cyst fluid (arrows indicate specific marker, in light blue are the molecules identified in both tissues). (c) MALDI MS profiles of 3 ovarian carcinomas vs benign tumor. (d) Hematoxilin eosin staining of the carcinoma slices before MALDI IMS for molecular localization of the ion at a m/z of 9744 (37 669 points with 300 shots by points at a 100 Hz repetition rate).

Alternate names: pa28(alpha); reg(alpha). Reg-alpha is a member of the proteasome activator 11S with a 249 amino acid residues protein (28 kDa). Ions at m/z 536.25, 700.68, 890.57, 970.47, 1158.55, and 1501.03 constitute trypsic fragments expected for this protein with 51% sequence coverage (Figure 2a,b). The principal m/z ions lead to the same identification for Reg-alpha trypsic fragments in nanoESI-q-TOF (Figure 2b,c). MALDI and ESI data allow getting 59 amino acid residues sequence from the 84 amino acid residues of the fragment (RIEDGNNFGA VQEKVFELMT SLHTKLEGFH TQISKYFSER GDAVTKAAKQ PHVGDYRQLV HELDEAEYRD IRLMVMEIRN AYAV) with a coverage percentage of 70% (Figure 2b). From the nanoESI spectra, some fragments have been validated by MS/ MS (CID) (Figure 2d) and confirm the identification. For example, the ion at m/z 501.26 is the MH33+ ion of the peptide QLVHELDEAEYR corresponding to a fragment of Reg-alpha. Taken together, the biomarker detected corresponds to the fragment (141-225) of Reg-alpha protein (Figure 2b). Sequence alignment of the 84 amino acid residues gave 100% identity with Reg-alpha (Q06323: accession Number, PSME1_Human), 63% with Reg-beta (Q9UL46: accession Number, PSME_human), and 46% with Reg-gamma (P61289: accession Number, PSME3_Human). Validation of Reg-Alpha as a Specific Ovarian Carcinoma Biomarker. Western blot analyses were performed on 9 carcinoma (6 adenocarcinoma, 1 endometrioid carcinoma, 2 mucinous carcinoma) as well as 16 benign tumors (1 mucinous cystadenoma, 2 serous cystadenoma, 2 serous cystadenofibroma, 6 endometriosic ovarian cysts, 3 teratoma, and 2 fibrothecoma). 4130

Journal of Proteome Research • Vol. 6, No. 11, 2007

The Reg-alpha antibody used was raised against the C-terminal Reg-alpha and showed the detection of the complete protein of 28 kDa in 88.8% of the carcinoma (6/9 cases) and 18.7% (3 slightly positives/16 cases) in benign tumors (Figure 2e) Immunocytochemical studies were performed on 11 carcinoma (8 adenocarcinoma, 1 endometrioid carcinoma, 1 mucinous carcinoma, and 1 neuroendocrine carcinoma) (Figure 3, pictures 1-8) versus 12 benign tumors (1 mucinous cystadenoma, 3 serous cystadenoma, 2 serous cystadenofibroma, 1 follicular ovarian cyst, 2 teratoma, and 3 fibrothecoma) (Figure 3, pictures 9-16). In 63.6% of carcinoma versus 16.6% benign tissues, labeling is observed. In all tissues, all the epithelial cells are labeled; however, epithelial cells are only very few in benign tissues compared to carcinoma, leading to the observation of much stronger labeling for carcinoma. Moreover, a cytoplasmic localization of Reg-alpha was found in carcinoma but never in benign nor in healthy tissues, whereas a nuclear localization was observed in 76.9% of benign tissues (Figure 4, pictures 1 and 2). Moreover, research in other gynecologic or extragynecologic epithelia revealed no labeling in normal ovaries (Figure 4, pictures 3 and 4) and a strong labeling in endocervical carcinoma (Figure 4, pictures 7 and 10). A weak labeling with nucleus localization was found in the case of the healthy endometria (Figure 4, picture 8) and fallopian tube (Figure 4, picture 9). In stomach adenocarcinoma (Figure 4, picture 12), the Reg-alpha labeling is high in contrast to the benign colon adenoma where the labeling is weak (Figure 4, picture 11). These results will be of great help for borderline cancer diagnosis (Figure 4, pictures 5 and 6).

MALDI Imaging/Profiling for Biomarker Hunting and Validation

research articles

Figure 2. (a) MALDI MS profile of the trypsin digestion of the peptide at a m/z of 9744. (b) Table of ions detected in MALDI (in red) and in ESI-q-TOF analysis after trypsin digestion and MS/MS analysis of the fragment. Sequence obtained after MALDI (red) and ESI-q-TOF (blue) fingerprint analyses. The coverage with Reg-alpha fragment sequence (141-225) is 70%. Also contains the complete sequence alignment of Reg-alpha. Amino acid residues in red are the ones obtained by MALDI and ESI-MS. (c) NanoESI-q-TOF profile of the trypsin digestion of the peptide at a m/z of 9744. (d) NanoESI-q-TOF MS/MS profile of the 5013+ ion precursor. The sequence obtained is QLVHELDEAEYR presenting 37% sequence identity with Reg-alpha fragment. (e) Western blot analyses with the anti-Ct Reg-alpha of the 16 benign tumors and 9 carcinomas.

To complete the validation, MALDI IMS (Figure 5a,f) studies were compared to ICC studies with the Reg-alpha C-term antibody (Figure 5d,f), specific MALDI IMS with tagged antibody (Figure 5e,f), and to hematoxilin coloration (Figure 5c,f). As it can be seen, the distribution of the Reg-alpha fragment is specific to epithelial cells and all three techniques corroborate each other.

Conclusion Taken together, we applied a mixed profiling and imaging strategy followed by targeted protein identification by protein micro-extraction and top-down tandem mass spectrometry to the study of carcinoma and benign ovarian tumors. Such a strategy has recently been used to investigate potential toxicity biomarkers in kidneys of rats that were administered gentamicin.30 In our case, 8 specific peptides have been identified in ovarian carcinoma, but we focused our attention on the most prevalent one, identified as a fragment of the 11S regulator complex (PA 28). The fragment of 84 amino acid residues is localized at the C-terminal part of the protein. The 11S regulator complex (syn.: PA28) is another regulatory complex associated with the 20S proteasome consisting of three subunits, alpha, beta, and gamma.31 Binding of 11S regulator complex to 20S proteasome does not depend on ATP hydrolysis and, unlike the 19S regulatory subunit, the 11S regulator complex does not catalyze degradation of large proteins but is

responsible for MHC-class l antigen processing;32-34 this is greatly improved by interferon gamma induced expression of alpha and beta subunits.35 A number of viral protein interactions with these proteasome subunits have been reported that may interfere with host anti-viral defense thereby contributing to mechanisms of cell transformation.36 The mechanisms by which it binds to the core particle via its subunits’ C-terminal tails and induces R-ring conformational changes to open the 20S gate suggest a similar mechanism for the 19S particle.31 The 20S proteasome has recently been detected in glioblastoma37 and in different cancer cell lines.35 However, no role in ovary cancer has yet been demonstrated for the 11S regulator complex. Our data are the first demonstrating the high level of expression of PA28 in carcinoma, especially in epithelial cells. Moreover, the distinct localization (nucleus or cytoplasm) of the C-terminal fragment of PA28 between benign and tumor cells could offer the possibility of a molecular diagnosis in the cases of borderline cancer. The detection at a high level of PA28 fragments in endocervical carcinoma and at a low level in no tumor gynecologic epithelia confirms its credibility as a potential cancer biomarker. Its presence in stomach adenocarcinoma epithelial cells and its absence in benign epithelial cells can be explained by the fact that PA28 can be expressed in solid tumor and fragments can perhaps be detected in human plasma as cancer indicator. PA28 activator belongs to the antigen processing machinery (APM). Its alteration by cleavage Journal of Proteome Research • Vol. 6, No. 11, 2007 4131

research articles

Lemaire et al.

Figure 3. Immunocytochemical data obtained after antigen retrieval technique and Hemalin eosin coloration with the anti-Ct Regalpha antibody. Pictures 1-8, Epithelial cells labeled with the Ct anti-Reg-alpha in ovarian carcinoma; picture 9, dermoide cyst; picture 10, dermoide cyst; picture 11, ovarian goiter; picture 12, serous cystadenofibroma; picture 13, serous cystadenoma; picture 14, serous cystadenoma; picture 15, follicular cyst; picture 16, fibrothecoma. (red arrows indicate the immunolabeling in benign tissue)

Figure 4. Immunocytochemical data obtained after antigen retrieval technique and H&S coloration with the anti-Ct Reg-alpha antibody. Picture 1, cytoplasmic localization of the anti-Ct Reg-alpha labeling in ovary carcinoma; picture 2, nuclear localization of the anti-Ct Reg-alpha labeling in ovary benign tumor; pictures 3 and 4, healthy ovary; pictures 5 and 6, immunolabeling with anti-Ct Reg-alpha in epithelial cells in borderline carcinoma; pictures 7 and 10, strong immunolabeling with anti-Ct Reg-alpha in endocervical carcinoma; pictures 8 and 9, nucleus mmunolabeling with anti-Ct Reg-alpha in heathy endometria and fallopian tube; pictures 11 and 12, strong immunolabeling with anti-Ct Reg-alpha in colon adenoma or stomach adenocarcinoma.

in ovarian carcinoma might be one mechanism to evade immune recognition. Such a hypothesis has already been advanced for the cases of APM chaperones as TAP, LMP2, 4132

Journal of Proteome Research • Vol. 6, No. 11, 2007

LMP10, and tapasin in the case of colon carcinoma, small cell lung carcinoma, and pancreatic carcinoma cell lines. In fact, IFN-γ treatment of these carcinoma cell lines corrected

research articles

MALDI Imaging/Profiling for Biomarker Hunting and Validation

Figure 5. Reg-alpha fragment distribution in ovarian carcinoma tissue. (a) Molecular MALDI imaging analysis of the Reg-alpha fragment distribution. (b) Ovarian carcinoma slice before hematoxilin coloration. (c) Ovarian carcinoma slice after hematoxilin eosin coloration. (d) Immunocytochemical studies with anti-Reg-alpha C-terminal antibody. (e) Specific MALDI imaging analysis with a primary antibody directed against the C-term part of human Reg-alpha and a secondary antibody modified with a photocleavage linker and a reporter peptide for MALDI MS detection. (f) Comparison on a zoomed portion of pictures presented, respectively, in panels e, c, and d showing correlation between the different approaches.

the TAP, LMP, and tapasin deficiencies and enhanced the PA28 R, LMP7, calnexin, and calreticulin expression which was then accompanied with increased levels of MHC class I antigens.38 Taken together, this study demonstrates for the first time a mixed profiling and imaging strategy followed by targeted protein identification by protein micro-extraction, top-down tandem mass spectrometry, and validation by specific MALDI imaging as well as immunocytochemistry which allow the detection and validation of pathological biomarkers. To definitely establish this potential biomarker, a multicentric study combining different hospitals is now necessary; this is now in progress.

Acknowledgment. Supported by grants from Centre National de la Recherche Scientifique (CNRS), Ministe`re de L’Education Nationale, de L’Enseignement Supe´rieur et de la Recherche (ACI Jeunes Chercheurs ACI JC4074 to I. Fournier), Institut National du Cancer (To I. Fournier and P. Ducoroy), Agence National de la Recherche (To I. Fournier), Projet International de Coope´ration Scientifique (to M. Salzet and R. Day), Conseil Re´gional Nord-Pas de Calais to M. Wisztorski. Also supported by a collaboration agreement between Bruker Daltonics GmbH and the Laboratoire de Neuroimmunologie des Anne´lides.

References (1) Edwards, B. K.; Brown, M. L.; Wingo, P. A.; Howe, H. L.; Ward, E.; Ries, L. A.; Schrag, D.; Jamison, P. M.; Jemal, A.; Wu, X. C.; Friedman, C.; Harlan, L.; Warren, J.; Anderson, R. N.; Pickle, L. W. Annual report to the nation on the status of cancer, 19752002, featuring population-based trends in cancer treatment. J. Natl. Cancer Inst. 2005, 97 (19), 1407-27. (2) Lambaudie, E.; Collinet, P.; Vinatier, D. [Ovarian cancers and CA 125 in 2006]. Gynecol. Obstet. Fertil. 2006, 34 (3), 254-7. (3) Ahmed, N.; Oliva, K. T.; Barker, G.; Hoffmann, P.; Reeve, S.; Smith, I. A.; Quinn, M. A.; Rice, G. E. Proteomic tracking of serum protein isoforms as screening biomarkers of ovarian cancer. Proteomics 2005, 5 (17), 4625-36. (4) Ardekani, A. M.; Liotta, L. A.; Petricoin, E. F., III. Clinical potential of proteomics in the diagnosis of ovarian cancer. Expert Rev. Mol. Diagn. 2002, 2 (4), 312-20. (5) Bandera, C. A.; Ye, B.; Mok, S. C. New technologies for the identification of markers for early detection of ovarian cancer. Curr. Opin. Obstet. Gynecol. 2003, 15 (1), 51-5. (6) Conrads, T. P.; Veenstra, T. D. The utility of proteomic patterns for the diagnosis of cancer. Curr. Drug Targets: Immune Endocr. Metab. Disord. 2004, 4 (1), 41-50. (7) Conrads, T. P.; Zhou, M.; Petricoin, E. F., III; Liotta, L.; Veenstra, T. D. Cancer diagnosis using proteomic patterns. Expert Rev. Mol. Diagn. 2003, 3 (4), 411-20. (8) Fields, M. M.; Chevlen, E. Ovarian cancer screening: a look at the evidence. Clin. J. Oncol. Nurs. 2006, 10 (1), 77-81.

Journal of Proteome Research • Vol. 6, No. 11, 2007 4133

research articles (9) Johann, D. J., Jr.; McGuigan, M. D.; Patel, A. R.; Tomov, S.; Ross, S.; Conrads, T. P.; Veenstra, T. D.; Fishman, D. A.; Whiteley, G. R.; Petricoin, E. F., III; Liotta, L. A. Clinical proteomics and biomarker discovery. Ann. N. Y. Acad. Sci. 2004, 1022, 295-305. (10) Kohn, E. C.; Mills, G. B.; Liotta, L. Promising directions for the diagnosis and management of gynecological cancers. Int. J. Gynaecol. Obstet. 2003, 83 (Suppl. 1), 203-9. (11) Lin, Y. W.; Lin, C. Y.; Lai, H. C.; Chiou, J. Y.; Chang, C. C.; Yu, M. H.; Chu, T. Y. Plasma proteomic pattern as biomarkers for ovarian cancer. Int. J. Gynecol. Cancer 2006, 16 (Suppl. 1), 139-46. (12) Liu, Y. Serum proteomic pattern analysis for early cancer detection. Technol. Cancer Res. Treat. 2006, 5 (1), 61-6. (13) Plebani, M. Proteomics: the next revolution in laboratory medicine? Clin. Chim. Acta 2005, 357, (2), 113-22. (14) Posadas, E. M.; Davidson, B.; Kohn, E. C. Proteomics and ovarian cancer: implications for diagnosis and treatment: a critical review of the recent literature. Curr. Opin. Oncol. 2004, 16 (5), 478-84. (15) Rapkiewicz, A. V.; Espina, V.; Petricoin, E. F., III; Liotta, L. A. Biomarkers of ovarian tumours. Eur. J. Cancer 2004, 40 (17), 2604-12. (16) Stevens, E. V.; Liotta, L. A.; Kohn, E. C. Proteomic analysis for early detection of ovarian cancer: a realistic approach? Int. J. Gynecol. Cancer 2003, 13 (Suppl. 2), 133-9. (17) Ahmed, N.; Barker, G.; Oliva, K. T.; Hoffmann, P.; Riley, C.; Reeve, S.; Smith, A. I.; Kemp, B. E.; Quinn, M. A.; Rice, G. E. Proteomicbased identification of haptoglobin-1 precursor as a novel circulating biomarker of ovarian cancer. Br. J. Cancer 2004, 91 (1), 129-40. (18) Bergen, H. R., III; Vasmatzis, G.; Cliby, W. A.; Johnson, K. L.; Oberg, A. L.; Muddiman, D. C. Discovery of ovarian cancer biomarkers in serum using NanoLC electrospray ionization TOF and FT-ICR mass spectrometry. Dis. Markers 2003, 19 (4-5), 23949. (19) Diamandis, E. P. Proteomic patterns in serum and identification of ovarian cancer. Lancet 2002, 360 (9327), 170; author reply 1701. (20) Engwegen, J. Y.; Gast, M. C.; Schellens, J. H.; Beijnen, J. H. Clinical proteomics: searching for better tumour markers with SELDITOF mass spectrometry. Trends Pharmacol. Sci. 2006, 27 (5), 251-9. (21) Fung, E. T.; Yip, T. T.; Lomas, L.; Wang, Z.; Yip, C.; Meng, X. Y.; Lin, S.; Zhang, F.; Zhang, Z.; Chan, D. W.; Weinberger, S. R. Classification of cancer types by measuring variants of host response proteins using SELDI serum assays. Int. J. Cancer 2005, 115 (5), 783-9. (22) Petricoin, E. F.; Ardekani, A. M.; Hitt, B. A.; Levine, P. J.; Fusaro, V. A.; Steinberg, S. M.; Mills, G. B.; Simone, C.; Fishman, D. A.; Kohn, E. C.; Liotta, L. A. Use of proteomic patterns in serum to identify ovarian cancer. Lancet 2002, 359 (9306), 572-7. (23) Wang, Q.; Li, L.; Li, D. R.; Zhang, W.; Wei, X.; Zhang, J. Q.; Tang, Y. [Identification of serum biomarkers for ovarian cancer using protein chips and time of flight mass spectrometry technology]. Zhonghua Fuchanke Zazhi 2006, 41 (8), 544-8. (24) Rai, A. J.; Zhang, Z.; Rosenzweig, J.; Shih Ie, M.; Pham, T.; Fung, E. T.; Sokoll, L. J.; Chan, D. W. Proteomic approaches to tumor marker discovery. Arch Pathol. Lab. Med. 2002, 126 (12), 151826.

4134

Journal of Proteome Research • Vol. 6, No. 11, 2007

Lemaire et al. (25) Xiao, Z.; Prieto, D.; Conrads, T. P.; Veenstra, T. D.; Issaq, H. J. Proteomic patterns: their potential for disease diagnosis. Mol. Cell. Endocrinol. 2005, 230 (1-2), 95-106. (26) Zhu, Y.; Wu, R.; Sangha, N.; Yoo, C.; Cho, K. R.; Shedden, K. A.; Katabuchi, H.; Lubman, D. M. Classifications of ovarian cancer tissues by proteomic patterns. Proteomics 2006, 6 (21), 5846-56. (27) Lemaire, R.; Tabet, J. C.; Ducoroy, P.; Hendra, J. B.; Salzet, M.; Fournier, I. Solid ionic matrixes for direct tissue analysis and MALDI imaging. Anal. Chem. 2006, 78 (3), 809-19. (28) Lemaire, R.; Wisztorski, M.; Desmons, A.; Tabet, J. C.; Day, R.; Salzet, M.; Fournier, I. MALDI-MS direct tissue analysis of proteins: Improving signal sensitivity using organic treatments. Anal. Chem. 2006, 78 (20), 7145-53. (29) Lemaire, R.; Stauber, J.; Wisztorski, M.; Van Camp, C.; Desmons, A.; Deschamps, M.; Proess, G.; Rudlof, I.; Woods, A. S.; Day, R.; Salzet, M.; Fournier, I. Tag-mass: specific molecular imaging of transcriptome and proteome by mass spectrometry base on photocleavable tag. J. Proteome Res. 2007, 6 (6), 2057-67. (30) Meistermann, H.; Norris, J. L.; Aerni, H. R.; Cornett, D. S.; Friedlein, A.; Erskine, A. R.; Augustin, A.; De Vera Mudry, M. C.; Ruepp, S.; Suter, L.; Langen, H.; Caprioli, R. M.; Ducret, A. Biomarker discovery by imaging mass spectrometry: transthyretin is a biomarker for gentamicin-induced nephrotoxicity in rat. Mol. Cell. Proteomics 2006, 5 (10), 1876-86. (31) Yang, Y.; Fruh, K.; Ahn, K.; Peterson, P. A. In vivo assembly of the proteasomal complexes, implications for antigen processing. J. Biol. Chem. 1995, 270 (46), 27687-94. (32) Kloetzel, P. M. The proteasome system: a neglected tool for improvement of novel therapeutic strategies? Gene Ther. 1998, 5 (10), 1297-8. (33) Rivett, A. J.; Gardner, R. C. Proteasome inhibitors: from in vitro uses to clinical trials. J. Pept. Sci. 2000, 6 (9), 478-88. (34) Rotem-Yehudar, R.; Groettrup, M.; Soza, A.; Kloetzel, P. M.; Ehrlich, R. LMP-associated proteolytic activities and TAP-dependent peptide transport for class 1 MHC molecules are suppressed in cell lines transformed by the highly oncogenic adenovirus 12. J. Exp. Med. 1996, 183 (2), 499-514. (35) Kuckelkorn, U.; Ferreira, E. A.; Drung, I.; Liewer, U.; Kloetzel, P. M.; Theobald, M. The effect of the interferon-gamma-inducible processing machinery on the generation of a naturally tumorassociated human cytotoxic T lymphocyte epitope within a wildtype and mutant p53 sequence context. Eur. J. Immunol. 2002, 32 (5), 1368-75. (36) Regad, T.; Chelbi-Alix, M. K. Role and fate of PML nuclear bodies in response to interferon and viral infections. Oncogene 2001, 20 (49), 7274-86. (37) Piccinini, M.; Rinaudo, M. T.; Anselmino, A.; Ramondetti, C.; Buccinna, B.; Fiano, V.; Ghimenti, C.; Schiffer, D. Characterization of the 20S proteasome in human glioblastomas. Anticancer Res. 2005, 25 (5), 3203-10. (38) Delp, K.; Momburg, F.; Hilmes, C.; Huber, C.; Seliger, B. Functional deficiencies of components of the MHC class I antigen pathway in human tumors of epithelial origin. Bone Marrow Transplant. 2000, 25 (Suppl. 2), S88-95.

PR0702722